English

Multimodal sensor fusion in the latent representation space

Artificial Intelligence 2025-05-29 v1 Human-Computer Interaction Machine Learning Signal Processing

Abstract

A new method for multimodal sensor fusion is introduced. The technique relies on a two-stage process. In the first stage, a multimodal generative model is constructed from unlabelled training data. In the second stage, the generative model serves as a reconstruction prior and the search manifold for the sensor fusion tasks. The method also handles cases where observations are accessed only via subsampling i.e. compressed sensing. We demonstrate the effectiveness and excellent performance on a range of multimodal fusion experiments such as multisensory classification, denoising, and recovery from subsampled observations.

Keywords

Cite

@article{arxiv.2208.02183,
  title  = {Multimodal sensor fusion in the latent representation space},
  author = {Robert J. Piechocki and Xiaoyang Wang and Mohammud J. Bocus},
  journal= {arXiv preprint arXiv:2208.02183},
  year   = {2025}
}

Comments

Under review for Nature Scientific Reports

R2 v1 2026-06-25T01:27:14.530Z